alianassmaaa/notification-timing-dataset

通知时间数据集包含100K合成样本，用于训练通知不良时间预测模型。数据集包含21个特征，涵盖时间上下文、电池状态、用户活动和通知历史。基于[C-3PO](https://arxiv.org/abs/1803.00458)的特征工程。数据集特征包括小时、电池水平、屏幕开启状态和不良时间等。数据集分为训练集、验证集和测试集，分别包含70000、15000和15000个样本。任务类别为表格分类，大小类别为10K<n<100K。

Notification Timing Dataset contains 100K synthetic samples for training notification bad-timing prediction models. 21 features covering time context, battery state, user activity, and notification history. Based on feature engineering from [C-3PO](https://arxiv.org/abs/1803.00458). Features include hour_of_day, battery_level, screen_on, and bad_timing. The dataset is divided into train, validation, and test sets with 70000, 15000, and 15000 samples respectively. Task category is tabular-classification, size category is 10K<n<100K.